sign in sign up
<<
Home , Dapedium, Ganoine

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/291973377

A preliminary study on the ornamentation patterns of ganoid scales in some Mesozoic actinopterygian

Article in Bollettino della Societa Paleontologica Italiana · December 2015 DOI: 10.4435/BSPI.2015.14

CITATIONS READS 0 206

2 authors:

Claudio Garbelli Andrea Tintori Nanging Institute of Geology and paleontolo… University of Milan

14 PUBLICATIONS 36 CITATIONS 118 PUBLICATIONS 1,487 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Middle fishes across the Tethys View project

All content following this page was uploaded by Andrea Tintori on 06 February 2016.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document and are linked to publications on ResearchGate, letting you access and read them immediately. TO L O N O G E I L C A A P

I '

T

A A

T

L

E

I

I

A Bollettino della Società Paleontologica Italiana, 54 (3), 2015, 219-228. Modena

C

N

O

A S

S. P. I.

A preliminary study on the ornamentation patterns of ganoid scales in some Mesozoic actinopterygian fishes

Claudio Garbelli & Andrea Tintori

C. Garbelli, Dipartimento di Scienze della Terra “Ardito Desio”, Università degli Studi di Milano, via Mangiagalli, 2, 20122 Milano, Italy; claudio. [email protected] A. Tintori, Dipartimento di Scienze della Terra “Ardito Desio”, Università degli Studi di Milano, via Mangiagalli, 2, 20122 Milano, Italy; andrea.tintori@ unimi.it

KEY WORDS - Ganoid scales, basal actinopterygians, ornamentation, squamation pattern.

ABSTRACT - Ganoid scales are one of the most common remains of vertebrates in the fossil record of Paleozoic and Mesozoic. Their knowledge is important for the understanding of the paleobiology and evolution of actinopterygian fishes. The distinctive feature of these scales is the outermost shiny layer of ganoine, an hyper-mineralized enamel-like tissue. During the Mesozoic, ganoid scales show a great variety of shapes and very different patterns of ornamentation made of ganoine. The surface of scales may be from smooth to very ornamented, with a relief arranged in tubercles, ridges, grooves or a combination of them. Here we present a preliminary study on the squamation of some Mesozoic non- actinopterygians in order to test the association between body shape and ganoine ornamentation. Using several morphological descriptors, we built an empirical morphospace to compare ganoid scales ornamentation. The use of a quantitative method to describe scales allows to test if there is a meaningful correlation between body shape and ganoine ornamentation in non-teleost actinopterygians. We found an important association between body shape and ornamentation in the taxa under investigation. In particular, deep-bodied fishes bear a more complex and variegate ornamentation than fusiform ones. This finding has important implications in our understanding of non-teleost actinopterygians paleobiology. Since swimming activity is a primary function for and this is performed by the body acting as an integrated unit, the correlation between body shape and ganoine ornamentation could be related to hydrodynamics. A quantitative test of this hypothesis, however, would be desirable.

RIASSUNTO - [Studio preliminare sugli schemi ornamentali delle scaglie ganoidi in alcuni pesci Attinopterigi del Mesozoico] - Le scaglie ganoidi sono uno dei più importanti e diffusi resti di vertebrati nel record fossile del Paleozoico e del Mesozoico. La loro conoscenza è importante per capire la paleobiologia e l’evoluzione dei pesci Attinopterigi. Il carattere distintivo di queste scaglie è la copertura esterna che è costituita dalla ganoina, un tessuto iper-mineralizzato simile allo smalto. Nel Mesozoico, le scaglie ganoidi sono estremamente differenziate e mostrano quindi una grande variabilità di forme e ornamentazioni prodotte dalla ganoina. La superficie delle scaglie può essere da liscia a fortemente ornata, con un micro-rilievo organizzato in tubercoli, creste, scanalature e dentellature, od una combinazione di essi. In questo lavoro abbiamo studiato le scaglie di alcuni Attinopterigi non-teleostei del Mesozoico, per verificare se esiste una relazione tra la forma del corpo e l’ornamentazione prodotta dalla ganoina. Utilizzando alcuni descrittori morfologici, abbiamo costruito un morfospazio empirico per confrontare le scaglie in termini di morfologia ed ornamentazione. L’uso di un metodo quantitativo per descrivere le scaglie ha permesso di verificare se c’è una correlazione significativa tra la forma del corpo e l’ornamentazione delle scaglie nei pesci Attinopterigii non-telostei. Abbiamo trovato un’associazione importante tra forma del corpo e descrittori morfologici delle scaglie nei casi studiati. In particolare, i pesci cosiddetti “deep-bodied” possiedono un’ornamentazione delle scaglie più complessa e differenziata di quella delle specie fusiformi. Questa osservazione ha delle importanti implicazioni per la nostra comprensione della paleobiologia dei pesci Attinopterigi non-teleostei. Infatti, siccome l’attività del nuoto è una funzione primaria per i pesci ed è attuata dall’organismo come se il corpo fosse un’unità integrata, la correlazione tra forma del corpo e ornamentazione della ganoina potrebbe essere relazionata all’idrodinamica. In futuro, un test quantitativo per verificare questa ipotesi sarebbe auspicabile.

INTRODUCTION potentially provide flexural constrains (Gemballa & Bartsch, 2002). The ganoid scales are a synapomorphic groundplan During the Paleozoic and Mesozoic and before the feature of ray-finned fishes, the (Schultze, vast teleostean radiation, ganoid scales were the common 1977; Patterson, 1982), and these structures showed a type of body covering among lower actinopterygians much greater amount of variability in deep time. Today, and their morphology was greatly variable, if compared only Polypteridae and Lepisosteidae retain ganoid scales with the scales of Polypteridae and Lepisosteidae (e.g., and their study allowed biologists and paleontologists Trinajstic, 1999a, b; Chen et al., 2012). This variability to clarify morpho-functional aspects of this kind of encompasses features such as the general shape of “exoskeleton”. Ganoid scales of these modern groups scales and the microrilief of ganoine covering, usually have several functional implications for swimming, such called “ornamentation”. Compared to the homogeneous as enhanced station-holding (Webb et al., 1992), control aspect of modern ganoid scales, primitive groups of of body stiffness and undulatory wave motion during actinopterygians show a much higher variability in terms steady swimming (Long et al., 1996); these are designed of scales shape and ornamentation of ganoine. The scales to permit extreme body curvatures in spite to the rigid from fossil record exhibit a rhomboid to very elongate structure of ganoid scales in the basal actinopterygian outline, from subcircular to squared. Their posterior fishes, structured with peg-and-socket articulation, which margin may be from smooth to serrated and their surface

ISSN 0375-7633 doi:10.4435/BSPI.2015.14 220 Bollettino della Società Paleontologica Italiana, 54 (3), 2015 may be from smooth to highly ornamented, with a a complete ganoid squamation were investigated: microsculpture composed of longitudinal or transversal ornatus Agassiz, 1843 (according to Tintori, ridges, isolated tubercles, crests and grooves. The meaning 1996); Gabanellia agilis Tintori & Lombardo, 1996; of these characters is not fully understood and some Stoppania gaetanii Lombardo et al., 2008; Felberia excelsa speculative hypotheses were proposed, mainly involving Lombardo & Tintori, 2004; Endennia licia Lombardo & functions related to swimming (Aleyev, 1977). One of the Brambillasca, 2005; Semiolepis brembanus Lombardo & most widespread ideas is that the surface roughness of fish Tintori, 2008; Dapedium politum Leach in De La Beche, skin may have some hydrodynamics functions related to 1822; Dapedium noricum Tintori, 1983; tomicus delaying the boundary layer separation or to the reduction Plieninger, 1847; Ptycholepis gracilis Davis, 1884; of skin friction drag (Fletcher et al., 2014). The boundary Ptycholepis bollensis Agassiz, 1833; Eugnathus philpotae layer separation can easily occur across smooth surfaces Agassiz, 1843; Luoxiongichthys hyperdorsalis Wen et in regions of adverse pressure gradient, but experimental al., 2012; Allolepidotus bellottii De Alessandri, 1910; measures suggest that there is no separation (Anderson Bobasatrania sp. White, 1932; Asialepidotus sp. Su, 1959. et al., 2001). The scales with ridges and grooves may Specimens belonging to these taxa were selected possibly promote the formation of microflows (Sudo et for study on the basis of quality preservation of their al., 2002). This could be consistent with the generation squamation pattern and scale morphology. The selected of a turbulent boundary layer in which the separation is specimens are deposited in the following museums: delayed, if compared to a laminar boundary layer. Some Museo di Paleontologia del Dipartimento di Scienze della authors argued that ganoine ornamentation could have Terra“ A. Desio”, Università degli Studi di Milano, Italy performed this function in extinct fishes (Burdak, 1986). (acronym MPUM); Museo Civico di Storia Naturale di In spite of the poorly known evolutionary significance Bergamo, Italy (MCSNB); Museo della Vicaria di S. of morphological pattern of ganoine, the study of Lorenzo in Zogno, Italy (MVSLZ); Museo Cantonale histology, crystallites arrangement and developmental di Storia Naturale di Lugano, Switzerland (MCSNL); aspects of the superficial hard tissues in the scales of Natural History Musuem of London, United Kingdom lower actinopterygians has received fair attention as (NHM); Geological Museum of the Peking University, potential phylogenetic character and its structure has been Paleontological collection, Beijing, China (GMPKU-P). described in detail for many fossil species (Schultze, 1966, When possible, holotype or paratypes were selected for 1977, 1992, 1996). These studies showed that ganoine the analysis. In other cases, specimens were selected on is a shiny, acellular, non-collagenous, hypermineralized the base of their preservation degree of ganoid squamation tissue of epidermal origin that covers the scales in and body features. The list of the studied specimens, with polypterids (), lepisosteids (), and a variety of their inventory numbers is in Table S1 (Supplementary other osteichthyans (Richter & Smith, 1995). It has been online material 1). observed that ganoine growth proceeds in concentric accretion and the tissue is arranged in multiple stratified Morphology of ganoid scales and ornamentation layers at maturity. The modulation in the accretion of Since the aim of this study is the investigation of the ganoine produces the so called “ganoine ornamentation”. ornamentation, the term “scales” is used herein to refer This ornamentation shows morphological convergences, as to the exposed portion of ganoid scales. The preservation, for example in the scales of certain and rich in details, of the external surface allows us to study Perleidiformes, which bear tubercles on the outer surface. the exposed portion of the scales in terms of ganoine In the evolutionary history of fishes, the presence of morphology, which is the component producing the convergent morphologies has been usually detected for ornamentation pattern. Three analogous, non-specialized various structures, such as fins or teeth (e.g., Donley et al., positions were selected along the body axis of the fishes 2004; Andreev, 2011). Since the factors that influence their to perform this study, in dorsal, lateral and caudal regions, external morphology are strongly related to swimming and respectively: positioning each of these regions, three hydrodynamics, a certain amount of correlation between scales have been selected. Scales from particular regions, morphological features is in fact expected, primarily due such as those close to the base of fins or in the anal region, to functional and structural constraints. were excluded because factors other than general body Thus, the aim of this study is to compare the pattern shape could affect their morphology (e.g., functionality of ganoine ornamentation in some ganoid fishes and to related to fin movements; adaptation to turbulences test if there is an association between the general body created by maneuvers during swimming performed using shape and the pattern of ganoine ornamentation. Specific fins). The overall morphological difference measured in aims of this work are: 1) to describe the ganoid scales analogous positions along the body axis is considered as in terms of ganoine morphology and ornamentation, 2) a tool to compare ornamentation patterns. to compare the different morphologies of scales using a In order to describe morphological variability of quantitative descriptive approach and 3) to verify if there scales, approaches based on landmarks are generally is a correlation between the body shape and the ganoine applied when there are anatomical homologous points to ornamentation pattern. use as references in the analysis (e.g., Chen et al., 2012). On the contrary, the morphological variation of irregular structures is more difficult to evaluate. This is the case of MATERIALS AND METHODS the ornamentation produced by ganoid scales. Here, we describe the ornamentation of ganoid scales using several Fourteen genera represented by sixteen different non-dimensional morphological descriptors, which do not species of Mesozoic non-teleost actinopterygians with require anatomical correspondence and are therefore useful C. Garbelli & A. Tintori - The ornamentation patterns of ganoid scales 221 to describe the organization of ganoine in terms of shape producing specific ornamentation. For each scale, a picture was acquired using a digital camera installed on a stereomicroscope. The outline of exposed surface of scales and of the ganoine microrelief were digitized. Measurements were taken on the binary image obtained from the digitalization of the scales: maximum diameter, height and length, perimeter and area. A schematization of measurements is shown in Fig. 1. Several non-dimensional parameters were subsequently calculated from the measurements to describe morphology (Russ, 1992). Fig. 1 - Example of measurements acquired for ganoid scales; Morphological descriptors used are: a) scale from lateral region of Paralepidotus ornatus under stereomicroscope; b) drawing of Fig. a with the reference for measures; the dashed line is the outline of the scale exposed; the Height light gray area is the surface covered by ganoine; Dmax = maximum Height-Length ratio (HLr) = diameter; H = height; L = length. Length

The Compactness is a numerical quantity representing Height-Length ratio is calculated on the exposed part the degree to which a shape is compact. It refers to ganoine of the scales; for the measurement of the height, starting coverage. The smaller the Compactness is, the more from the middle of the ventral edge of the scale, we traced complex the ganoine pattern is. a line parallel to the anterior edge, ending to the dorsal edge. For the measurement of the length, starting from Covered Area the middle of the anterior edge, we traced a line parallel Cover - coefficient = to the ventral edge, ending to the posterior edge. Exposed Area

The Cover-coefficient is the ratio between the area 4π * Area Formfactor = 2 covered by ganoine and the total area of scale exposed. Perimeter It shows a positive correlation with the ganoine coverage. This parameter may be also indicative of the relative lightening of scale coverage. The Formfactor is the inverse ratio of the squared To compare the morphologies of different ganoid perimeter of an object to the squared perimeter of a scales, we performed a Principal Component Analysis circle of the same surface. The area and the perimeter using the standardized values of morphological descriptors. are calculated on the ganoine outline. The smaller Since the morphological descriptors are partially related to the Formfactor is, the more indented the outline is. In each other, it is necessary to plot a new set of variables (the particular, since the main variability of scales outline Principal Components), which are orthogonal and represent is on posterior margin, which may be from smooth to an empirical and parametric morphospace. In this space it serrate, the Formfactor provides information about the is possible to compare morphologies using the Euclidean degree of serration. metric. Subsequently, we performed a Hierarchical Cluster Analysis using Euclidean metric distances to group scales on the basis of morphology. In order to choose the best number of groups for subdividing the scales, we performed 4 * Area a Cluster Validation analysis (Internal Validation using Roundness = 2 π * MaxDiameter Dunn, Silhouette and Collectivity test; Brock et al., 2008). All these analyses are performed using R 3.2.2 (R Development Core Team, 2015).

The Roundness is the ratio between the actual area and Body morphology vs. pattern of ganoine ornamentation the area of a circle of the same diameter: the higher the To establish the existence of a significant correlation Roundness is, the more rounded the scale surface shape between ganoine ornamentation and overall morphology is. The area and maximum diameter are calculated on of the taxa under investigation, the body shape was ganoine outline because the use of this parameter is aimed captured using the geometric morphometric approach to describe the ornamentation morphology. (Zelditch et al., 2004), performing two different analyses. In the first one, homologous landmarks were selected on the basis of their functional or ecologic role following √(4/π Area) studies about the shape variation in extinct or modern Compactness = * fishes (Fig. S1). In the second analysis, 80 semilandmarks, MaxDiameter equally spaced and aligned using the minimum bending energy criterion, were used to define the body outline. 222 Bollettino della Società Paleontologica Italiana, 54 (3), 2015

The landmarks and semilandmarks were acquired and digitized using TPSDIG, version 2.17 (Rohlf, 2013a). Subsequently shape variables were extracted by applying Generalized Procrustes Analysis (GPA; Rohlf & Slice, 1990). This algorithm aligns landmarks configurations to a common reference (the consensus) after removing the effect of rotation, translation, and differences in size among specimens. Partial Least Square analysis (PLS; Rohlf & Corti, 2000) was applied to detect a possible association between shape variables and morphological descriptors of scales. The Partial Least Square analysis (PLS) extracts pairs of orthogonal latent variables from the correlation matrices of each block, using singular value decomposition. Each pair of latent variables (Singular Warps; SW) maximizes the possible co-variation between the two blocks. In PLS models there are no predictors and predicted variables, but both blocks of variables (shape variables and morphological descriptors of scales) are equally weighted. A nonparametric permutation test Fig. 2 - Cross plot of the first two Principal Components (PC), (with 999 random repetitions) was performed in order to resulted from the analysis of morphological descriptors of scales. test if the covariation of latent variables and correlation The groups are obtained from the Hierarchical Cluster Analysis. coefficients were significant. GPA, PLS analyses and From group A to C, there is an increment of the aspect ratio of permutation test were performed using TPSRELW and scales and the ganoine pattern becomes more and more elaborated. TPSPLS, version 1.19 (Rohlf, 2010, 2013b).

variance (Tab. 1). The Cluster Validation analysis highlights RESULTS that the best scores are reached using a hierarchical cluster method. The number of clusters with the best score is 2 for Measures, shape descriptors and scales classification Connectivity and Silhouette test, but it is grater for Dunn The acquired measurements are summarized in the (see Supplementary online material 1: Tab. S2, Fig. S3). Supplementary online material 2. The largest scales exceed Looking the results of cluster validation and the characters 50 mm2 in exposed surface, and most of them belongs of scales, we evaluated that a good compromise for the best to the lateral body regions of large deep-bodied fishes grouping number is three (Fig. 2). The PCA reveals that such as Sargodon tomicus and Stoppania gaetanii. The along the first PC axis there is a decrease of Formfactor, smallest scales belong to the paleonisciforms Ptycholepis Circularity, Compactness and Cover-coefficient, and an and Bobasatrania, reaching less than 1 mm2 in exposed increase in the Height-Length ratio of the scales. The surface. A summary of the shape descriptors ranges is clusters follow the trend along the first PC axis. presented in the Supplementary online material 1 (Fig. The resulted clusters include these types of scales: S2). Roundness and compactness show a wide range of overlapping values for the scales in the three different body cluster A: most of the caudal scales belongs to this positions of the investigated specimens. The Formfactor of group; they bear a complete cover of ganoine and have ganoine coverage is more differentiated for the dorsal and an aspect ratio close to 1, with a diamond shape (Fig. lateral scales, compared to caudal ones, which present the 3a). Variegate lateral and dorsal scales are included in highest values of this parameter. The Height-Length ratio this cluster, both those with an elongated aspect ratio, but shows lower values in caudal scales than in other body with a smooth and complete cover of ganoine (Fig. 3b), positions. The Cover-coefficient has a wide range of values and those with a sub-quadrate outline and a reduction both in lateral and dorsal scales, whereas in most of the of ganoine cover (Fig. 3c). The scales of this group are caudal ones, this parameter is equal to 1. characterized by weak to developed ganoine serrations The Principal Component Analysis of the non- along the posterior edge, which vary in number and depth dimensional shape descriptors results in the first two PCs, of the elements. explaining more than the 82% of the observed variance, with the first PC explaining about 68% of the observed cluster B: most of the scales exhibits an aspect ratio very different from 1; the scales are partially ganoine free in between the elements composing the ornamentation; Principal components scales strongly covered by ganoine show well developed PC1 PC2 PC3 PC4 PC5 and deep serrations with high number of elements (Fig. 3d); scales ornamented with longitudinal grooves tend to SD 1.835 0.854 0.773 0.484 0.188 have a lower aspect ratio and show reduced coverage of Proportion of 0.679 0.147 0.120 0.047 0.007 ganoine in between the ridges (Fig. 3e). variance cluster C: this group contains scales ornamented by Tab. 1 - Summary of the results of Principal Component Analysis performed on the standardized values of morphological descriptors isolated elements of ganoine on the surface of scales; these used to compare ganoid scales. scales have an high aspect ratio (Fig. 3f). C. Garbelli & A. Tintori - The ornamentation patterns of ganoid scales 223

Fig. 3 - Ganoid scales belonging to different clusters. a) Caudal scale of Paralepidotus ornatus (specimen MVSLZ ST82916) showing diamond shape, with a complete coverage of smooth ganoine, scale bar: 1 mm; b) lateral scales of Semiolepis brembanus (MPUM 9288), which is slightly higher than longer, completely covered by smooth ganoine and with thin serration in the posterior margin, scale bar: 1 mm; c) dorsal scale of Paralepidotus ornatus (MVSLZ ST82916) showing a sub-quadrate outline and a partial reduction of the ganoine coverage producing deep serration, scale bar: 3 mm; d) lateral scales of Asialepidotus sp. (GMPKU-P 3025), higher than longer and with deep regular serration in the posterior margin, scale bar: 4 mm; e) lateral scales of Ptycholepis bollensis (NHM P858a), longer than higher with an ornamentation of longitudinal grooves; in between the grooves, the ganoine is thin or absent, scale bar: 1 mm; f) lateral scales of Stoppania gaetanii (MPUM 9542), with high HLr and the ganoine showing a complex ornamentation producing tubercles, scale bar: 2.5 mm.

Correlation between body shape and ornamentation DISCUSSION PLS analysis supports the presence of an association between the morphological descriptors of scales and Scale morphology and ornamentation of ganoine. body shape in the actinopterygians analyzed. For the The external morphology of the studied scales is analysis in which landmarks were used to define body varied, both in scale shape and in ganoine coverage. We shape, the first pair of SW vectors explains the 82.15% have included in the analysis: of the covariance and their correlation coefficient is equal to 0.758. The permutation tests are significant for the 1. diamond shaped scales, totally covered by ganoine, correlation of vectors, but not for the amount of variance with a smooth posterior margin (Fig. 3a); explained. For the analysis in which semilandmarks (i.e., 2. diamond shaped scales, totally covered by ganoine, outline) were used, the first pair of SW vectors explains with a serrated posterior margin (fig. 6C in Lombardo et the 93.73% of the covariance and their correlation al., 2008); coefficient is equal to 0.842. The permutation tests are 3. scales with HLr major than 1, totally covered by significant both for the amount of variance explained by ganoine, with a thin serrated posterior margin (Fig. 3b); the first SW vectors pair and the correlation of vectors. 4. scales with HLr major than 1, not totally covered by A summary of the other latent variables is provided in ganoine, with deep serrations along the posterior margin Tab. 2. (Fig. 3d); 224 Bollettino della Società Paleontologica Italiana, 54 (3), 2015

Explained Significativity of covariance Significativity of correlation SW Covariance Correlation covariance (999 permutations) (999 permutations)

1 0.251 82.15 0.758 p ≤ 0.11 p ≤ 0.03* 2 0.086 9.64 0.785 p ≤ 0.89 p ≤ 0.04* Landmarks 3 0.071 6.63 0.717 p ≤ 0.53 p ≤ 0.06 4 0.024 0.76 0.663 p ≤ 0.99 p ≤ 0.19 5 0.016 0.34 0.766 p ≤ 0.99 p ≤ 0.06 1 0.334 93.73 0.842 p ≤ 0.046* p ≤ 0.001* 2 0.070 4.09 0.726 p ≤ 0.938 p ≤ 0.053 Outlines 3 0.043 1.56 0.798 p ≤ 0.901 p ≤ 0.006* 4 0.023 0.45 0.542 p ≤ 0.932 p ≤ 0.293 5 0.010 0.08 0.702 p ≤ 0.994 p ≤ 0.045*

Tab. 2 - Summary of the results obtained from PLS analysis between body shape block and morphological descriptors of scales data block for the 16 taxa of Mesozoic actinopterygians under investigation. The covariance and correlation values and the permutation test results are represented for the first five Singular Warps.

5. scales with HLr minor than 1, with longitudinal since scales with different types of ornamentation are ridges of ganoine, with free grooves in between (Fig. 3e); clustered together (Fig. 2). The cluster A is the most 6. scales with HLr much more than 1, partially covered inconsistent, instead the B and C are progressively more by ganoine, which tend to form isolated tubercles on the and more consistent. This could be related to two factors: outer surfaces (Fig. 3f). 1. the morphology of scales changes gradually along the body of fishes of a specific taxon and, as a consequence, The clusters of scales based on shape descriptors are it results as a continuously variable character, difficult to not totally consistent with the observed morphologies, classify and clusterize;

Fig. 4 - Cross plot of the first pair of Singular Warps (SW) representing body shape block (x-axis, a: body landmarks, b: body outlines) and morphological descriptors data block (y-axis) for 16 species of Mesozoic actinopterygians. 1) Allolepidotus bellottii (4431), 2) Asialepidotus sp. (GMPKU-P 3025), 3) Bobasatrania sp. (coll. Tintori), 4) Dapedium noricum (MCSNB 3316), 5) Dapedium politum (NHM P1585), 6) Endennia licia (MPUM 8434), 7) Eugnathus philpotae (NHM P3576), 8) Felberia excelsa (MCSNL 5034), 9) Gabanellia agilis (MPUM 7751), 10) Luoxiongichthys hyperdorsalis (GMPKU-P 1580), 11) Paralepidotus ornatus (MVSLZ ST82916), 12) Ptycholepis bollensis (NHM P858a), 13) Ptycholepis gracilis (NHM P7789), 14) Sargodon tomicus (MPUM 7516), 15) Semiolepis brembanus (MPUM 9288), 16) Stoppania gaetanii (MPUM 9542). C. Garbelli & A. Tintori - The ornamentation patterns of ganoid scales 225

dorsal region of S. gaetanii, and this explains the position of the occupied area.

Relationships between body shape and morphological descriptors of the scales The PLS analysis reveals that the overall scales ornamentation is better associated to the outline rather than shape defined by landmarks (Tab. 2, Fig. 4). Fishes with similar body shape show more similar trend for overall ornamentation pattern. This highlights that, even if the single scales are very different from each other, the pattern of ornamentation can be considered similar. In fact, fishes that have closer positions in the PLS (Fig. 4), possess similar organization of the ornamentation, irrespectively Fig. 5 - Cross plot of the first two Principal Components (PC) for to the complexity of the ganoine organization on scale the shape descriptors, which refer to the morphologies of scales surfaces. For examples, Gabanellia agilis and Ptycholepis represented in Fig. 3. The letters refer to Fig. 3; the number on the sp. are very close (Fig. 4, G. agilis = 9, Ptycholepis sp.= black arrows represent Euclidean distances between the caudal scale 12/13), sharing the streamlined outline of the body. of Paralepidotus ornatus (a), the lateral one of Allolepidotus bellottii (b) and the lateral one of Stoppania gaetanii (f). G. agilis has scales totally covered by a thin layer of ganoine and with a weakly serrated posterior margin (Tintori & Lombardo, 1996). Instead, Ptycolephis sp. 2. the shape descriptors are not fully exhaustive and has thick ganoid scales, with longitudinal grooves along some features that contribute to morphological variability all the body (see Fig. 3e for lateral scale). For both taxa, may be underestimated or not accounted for (e.g., the the main changes are related to the aspect ratio of the thickness of the ganoine coverage). scales, but the organization of ganoine ornamentation Despite the difficulty of creating consistent clusters, remains costant. Deep bodied fishes, such as Stoppania the PCA analysis of morphological descriptors represents gaetanii and Felberia excelsa, have a more differentiated a useful tool to compare scales, since it generates an ornamentation, with a progressive change in the empirical and parametric morphospace, where it is ornamentation of the scales along the body (Lombardo possible to quantify differences between scales. For & Tintori, 2004; Lombardo et al., 2008). Taxa such as the scales under investigation, we observed that cluster Paralepidotus ornatus and Dapedium sp. have a certain A is closer to cluster B than to C. This means that the amount of change in scales ornamentation, with the dorsal ornamentation of a scale in A is more similar to a scale and lateral being more ornamented and the caudal being in B than in C (Fig. 2). For example, a lateral scale of smooth. This results in a intermediate position of these Asialepidotus sp. (Fig. 3b) can be evaluated to be more taxa (Fig. 4, Dapedium sp. = 4/5, P. ornatus = 11). similar to a caudal one of Paralepidotus ornatus (Fig. 3a) than to a lateral scale of Stoppania gaetanii (Fig. Implications 3f) in terms of the organization pattern of ganoine. The relationship between body shape and ganoine This is in agreement with the morphospace occupancy ornamentation could be potentially extended to other plotted in Fig. 5, where it is possible to measure metric non-teleost actinopterygians not examined in this study. distances between scales. As a consequence, the overall morphospace occupied by the scales of a single taxon can be considered an index of the differentiation of scales morphology. This allows to compare in a quantitative way the ornamentation pattern of different taxa. For example, Gabanellia agilis shows a relative homogeneous type of scales in the body positions analyzed. The scales are totally covered by ganoine and present posterior serrations. The main variability is mostly related to the change in aspect ratio of the scales. Instead, S. gaetanii has scales that gradually switch from high aspect ratio, with ganoine tuberculation, to smooth scales, with sub-squared outline, completely covered by ganoine. This results in a different morphospace occupancy for the scales of each taxon (Fig. 6), with S. gaetanii occupying a greater area because of the more differentiated scales morphology with respect to G. agilis. In addition, the position of the occupied area is informative about the type of scales ornamentation. In fact, it is noteworthy that Bobabsatrania sp., which has homogeneous scales ornamentation composed of Fig. 6 - Cross plot of the first two Principal Components (PC) for elongated ridges on scales surface, shows a morphospace the shape descriptors referring to the scales of Stoppania gaetanii, area similar to G. agilis. On the other hand, Bobasatrania Gabanellia agilis and Bobasatrania sp.; the numbers represent the sp. has a type of ornamentation more similar to the lateral/ area of the convex hulls for each taxon represented. 226 Bollettino della Società Paleontologica Italiana, 54 (3), 2015

For example, the deep-bodied Kyphosichthys grandei (Xu 1983). However, complex ornamentations, such as &Wu, 2012) possesses an ornamentation that varies along random grooves or tuberculations, are difficult to relate its body, as observed also in specimens of Dapedium from to a specific function. the Lower of the Holzmaden area (Thies & Hauff, 2011). On the other hand, other fusiform taxa, with ganoid scales, have a relatively homogeneous CONCLUSIONS ornamentation, such as Isanichthys palustris among the Semionotiformes or Perleidus among the Perleidiformes In this study, we have highlighted that the ganoid (Cavin & Suteethorn, 2006; Lombardo et al., 2011). scales and their ornamentation can be studied using In this study, we have included representatives a comparative and quantitative approach. We have of different orders, which show distinctive traits of performed a preliminary study to solve the thorny ornamentation. The paleonisciforms, Bobasatrania sp. question, which is the interpretation of structures that and Ptycholepis sp., show a pronounced and complex today are present in few taxa, but that were very common ornamentation of scales, made of ridges and grooves, as in Paleozoic and Mesozoic fish lineages, such as the observed in other genera of basal actinopterygians (see ganoid scales, with their peculiar ornamentation, which for instance the genus Challaia in López-Arbarello et is no longer observed in extant lineages. Applying this al., 2010). method, we have demonstrated that there is a general The deep bodied perleidiforms Felberia excelsa association between the body shape and the ornamentation and Stoppania gaetanii have a complex ornamentation pattern of ganoine in the investigated taxa, highlighting too, which changes drastically in the different positions a possible function related to the hydrodynamics of analyzed. On the other hand, the fusiform Gabanellia swimming. agilis shows smooth scales ornamented only with This finding is very important to better understand posterior serrations. The Perleidiformes seem to bear a basal actinopterygians adaptations and paleobiology, variety of types of ornamentation, as for example the since the ganoid scales are the most usual remains of genus Colobodus, which shows a very peculiar pattern of these organisms. A quantitative test of our hypothesis longitudinal ridges and grooves (Rusconi et al., 2006; see would be desirable and the approach used in this study fig. 10 in Sun et al., 2008). The slender semionotiforms should be extended to other fossil materials and to other tend to have smooth scales ornamented with posterior time intervals. serration. In the taxa developing a prominent hump-back region, such as Paralepidotus ornatus, the scales of this region bear tubercles (see for instance Scheenstia zappi in ACKNOWLEDGEMENTS López-Arbarello & Sferco, 2011). Otherwise, deep-bodied We thank Emma Bernard for the support during the visit to taxa posses scales of dorsal and lateral regions completely Natural History Museum of London, Anna Paganoni for having covered by isolated tubercles (see Sargodon tomicus). provided the access to the specimens of the Museo Civico di Storia Even if there are differences in the specific types Naturale di Bergamo and Cristina Lombardo for the continuous of ornamentation of different taxa, it is clear that there support during the study of specimens housed in the Museo del is a general association between ornamentation pattern Dipartimento di Scienze della Terra “A. Desio”. Thanks to the of ganoid scales and the body shape. One possibility critical review of the manuscript provided by Giorgio Carnevale is that the observed trend could have some adaptive and Giuseppe Marrammà. This paper has greatly benefited of the meanings. During the evolutionary history of fishes, financial support of the Università degli Studi di Milano (funds repeated morphological convergences occurred to improve provided by Lucia Angiolini). The BSPI Editorial Board is also acknowledged. swimming performances (Fletcher et al., 2014). Since fishes work as an integrated unit and the association of morphological characters is expected to produce better REFERENCES performances, we can hypothesize that the ornamentation is related to the hydrodynamic forces shaping the Agassiz L. (1833–1843). Recherches sur les poissons fossils, 5 general morphology of fishes. Several authors suggest vols. cvii + 1798 pp. Petitpierre, Neuchâtel et Soleure, France. that the microrelief of scales has some functions related Aleyev Y.G. (1977). Nekton. vi + 435 pp. Dr. W. Junk, The Hague. to the control of water flow around the fish body (drag Anderson E.J., McGillis W.R. & Grosenbaugh M.A. (2001). The boundary layer of swimming fish. Journal of Experimental reduction, control of the turbulence, delay of boundary Biology, 204: 81-102. layer detachment, enhancement of float; e.g., Bone, Andreev P. (2011). Convergence in dental histology between the 1972; Lang et al., 2008). This has been proven in part for Late Triassic semionotiform Sargodon tomicus () placoid scales of (Oeffner & Lauder, 2012), but, at and a Late (Turonian) pycnodontid (Neopterygii: present, only conjectures have been provided for ganoid Pycnodontiformes) species. Microscopy Research and scales (Burdak, 1986). The hydrodynamic function has to Technique, 74: 464-479. be verified, since ganoine produces structures that could Bone Q. (1972). Buoyancy and Hydrodynamic Functions of have other roles. For example, Dipteronotus olgiattii Integument in the Castor Oil Fish, Ruvettus pretiosus (Pisces: (Tintori, 1990) possesses scales in which ganoine produces Gempylidae). Copeia, 1: 78-87. Brock G., Datta S., Pihur V. & Datta S. (2008). clValid: an R package prominent spines, which could have a protection function. for cluster validation. Journal of Statistical Software, 25: 1-22. In addition, structural functions cannot be excluded. For Burdak V.D. (1986). Morphologie fonctionnelle du tégument example, it has been suggested that the ridges on the écailleux des poissons. Cybium, 10 (Suppl.). 147 pp. scales of Bobasatrania contribute to the body stiffness Campbell K.S.W. & Duy Phuoc L. (1983). A Late necessary for the swimming (Campbell & Duy Phuoc, actinopterygian fish from Australia.Palaeontology , 26: 33-70. C. Garbelli & A. Tintori - The ornamentation patterns of ganoid scales 227

Cavin L. & Suteethorn V. (2006). A new semionotiform Richter M. & Smith M.M. (1995). A microstructural study of the (Actinopterygii, Neopterygii) from Upper Jurassic-Lower ganoine tissue of selected lower vertebrates. Zoological Journal Cretaceous deposits of north-eastern Thailand, with comments on of the Linnean Society, 114: 173-212. the relationships of semionotiforms. Palaeontology, 49: 339-353. Rohlf F.J. (2010). TpsRelw, relative warps analysis, version 1.49. Chen D., Janvier P., Ahlberg P.E. & Blom H. (2012). Scale Department of Ecology and Evolution, State University of New morphology and squamation in the osteichthyan York, Stony Brook, New York. from Gotland, Sweden. Historical Biology, 24: 411-423. Rohlf F.J. (2013a). tpsDig 2.17. Department of Ecology and Davis J.W. (1884). Description of a new species of Ptycholepis Evolution, State University of New York, Stony Brook, New from the Lias of . Annals And Magazine of Natural York. History, 13: 335-337. Rohlf F.J. (2013b). tpsPLS, Version 1.19. Department of Ecology De Alessandri G. (1910). Studi sui Pesci Triasici della Lombardia. and Evolution, State University of New York, Stony Brook, Memorie della Società Italiana di Scienze Naturali, 7: 1-147. New York. De La Beche H.T. (1822). Remarks on the Geology of the Rohlf F.J. & Corti M. (2000). Use of two-block partial least squares South Coast of England, from Bridport Harbour, , to to study covariation in shape. Systematic Biology, 49: 740-753. Babbacombe Bay, Devon. Transactions of the Geological Rohlf F.J. & Slice D.E. (1990). Extensions of the Procrustes method Society of London, Series 2, Volume 1: 40-47. for the optimal superimposition of landmarks. Systematic Donley J.M., Chugey A.S., Konstantinidis P., Gemballa S. & Zoology, 39: 40-59. Shadwick R.E. (2004). Convergent evolution in mechanical Rusconi M.R., Lombardo C. & Tintori A. (2006). Colobodontide design of lamnid sharks and . Nature, 429: 61-65. fro the Upper Triassic (Carnian) of Friuli Venezia Giulia (Udine Fletcher T., Altringham J., Peakall J., Wignall P. &. Dorrell R. Italy). Gortania, Atti Museo Friulano di Storia Naturale, 28: (2014). Hydrodynamics of fossil fishes. Proceedings of the 59-72. Royal Society B-Biological Sciences, 281: 20140703. Russ J.C. (1992). The Image Processing Handbook. 445 pp. CRC Gemballa S. & Bartsch P. (2002). Architecture of the integument Press Inc., Boca Raton (Florida). in lower teleostomes: Functional morphology and evolutionary Schultze H.P. (1966). Morphologische und histologische implications. Journal of Morphology, 253: 290-309. Untersuchungen an Schuppen mesozoischer Actinopterygier Lang A.W., Motta P., Hidalgo P. & Westcott M. (2008). Bristled (Übergang von Ganoid-zu Rundshuppen). Neues Jahrbuch für skin: a microgeometry for boundary layer control? Geologie und Paläontologie, 126: 232-314. Bioinspiration & Biomimeics, 3: 046005. Schultze H.P. (1977). Ausgangsform und Entwicklung Lombardo C. & Brambillasca F. (2005). A New perleidiform der rhombischen Schuppen der (Pisces). (Actinopterygii, Osteichthyes) from the Late Triassic of Paläontologisches Zeitschrift, 51: 152-168. Northern Italy. Bollettino della Società Paleontologica Italiana, Schultze H.P. (1992). Early actinopterygians (Osteichthyes, 44: 25-34. Pisces) from Siberia. In Mark-Kurik E. (ed.), Fossil fishes as Lombardo C., Rusconi M. & Tintori A. (2008). New Perleidiform living , 1: 233-242, Tallinn, Estonia: Academia. from the Lower Ladinian (Middle Triassic) of the Northern Schultze H.P. (1996). The scales of Mesozoic Actinopterygians. In Grigna (Northern Italy). Rivista Italiana di Paleontologia e Arratia G. & Vioehl G. (eds), Mesozoic fishes 1 - Systematics Stratigrafia, 114: 263-272. and Paleoecology: 83-93, Verlag F. Pfeil, Munchen. Lombardo C., Sun Z.Y., Tintori A., Jiang D.Y. & Hao W.C. Su T. (1959). Triassic Fishes from Kueichow, South-West China. (2011). A new species of the genus Perleidus (Actinopterygii: Vertebrata PalAsiatica, 3: 205-215. Perleidiformes) from the Middle Triassic of Southern China. Sudo S., Tsuyuki K., Ito Y. & Ikohagi T. (2002). A study on the Bollettino della Società Paleontologica Italiana, 50: 75-83. surface shape of fish scales.JSME international journal. Series Lombardo C. & Tintori A. (2004). New perleidiformes from the C, Mechanical Systems, Machine Elements and Manufacturing, Triassic of the Southern Alps and the revision of Serrolepis from 45: 1100-1105. the Triassic of Württemberg (Germany). Mesozoic Fishes 3 - Sun Z.Y., Tintori A., Lombardo C., Jiang D.Y., Hao W.C., Sun Y.L., Systematics, Paleoenvironments and Biodiversity, 3: 179-196. Wu F.X. & Rusconi M. (2008). A new species of the genus Lombardo C. & Tintori A. (2008). A new semionotid fish Colobodus Agassiz, 1844 (Osteichithyes, Actinoperygii) from (Actinopterygii, Osteichthyes) from the Late Triassic of the the Pelsonian (Anisian, Middle Triassic) of Guizhou, South Northern Italy. In Arratia G., Schultze H.-P., Wilson M.V.H., China. Rivista Italiana di Paleontologia e Stratigrafia, 114: (eds), Mesozoic Fishes 4 - and phylogeny: 129-142, 363-376. Verlag F. Pfeil, Munchen. Thies D. & Hauff R.B. (2011). New species of Dapedium LEACH, Long J.H., Hale M.E., McHenry M.J. & Westneat M.W. (1996). 1822 (Actinopterygii, Neopterygii, Semionotiformes) from the Functions of fish skin: flexural stiffness and steady swimming Early Jurassic of South Germany. Palaeodiversity, 4: 185-221. of longnose osseus. The Journal of Experimental Tintori A. (1983). Hypsisomatic Semionotidae (Pisces, Biology, 199: 2139-2151. Actinopterygii) from the Upper Triassic of Lombardy (N. Italy). López-Arbarello A., Rauhut O.W.R. & Cerdeño E. (2010). The Rivista Italiana di Paleontologia e Stratigrafia, 88: 417-442. Triassic fish faunas of the Cuyana Basin, Western Argentina. Tintori A. (1990). Dipteronotus olgiatii n. sp. (Actinopterygii, Palaeontology, 53: 249-276. Perleidiformes) from the Kalkschieferzone of Ca’ del Frate (n. López-Arbarello A. & Sferco E. (2011). New semionotiform Italy) (preliminary note). Atti Ticinesi di Scienze della Terra, (Actinopterygii: Neopterygii) from the Late Jurassic of southern 33: 191-19. Germany. Journal of Systematic Palaeontology, 9: 197-215. Tintori A. (1996). Paralepidotus ornatus (Agassiz 1833-43): A Oeffner J. & Lauder G.V. (2012). The hydrodynamic function of semionotid from the (Late Triassic) of Europe. In Arratia shark skin and two biomimetic applications. The Journal of G. & Vioehl G. (eds), Mesozoic fishes 1 - Systematics and Experimental Biology, 215: 785-795. Paleoecology: 167-179, Verlag F. Pfeil, Munchen. Patterson C. (1982). Morphology and Interrelationships of Primitive Tintori A. & Lombardo C. (1996). Gabanellia agilis gen. n. sp. n. Actinopterygian Fishes. American Zoologist, 22: 241-259. (Actinopterygii, Perleidiformes) from the Plieninger W.H.T. (1847). Microlestes antiquus und Sargodon of Lombardy (North Italy). Rivista Italiana di Paleontologia e tomicus inder Grenzbreccia des Keupers von Degerloch. Jahresh Stratigrafia, 102: 227-236. Vel vaterl Naturk Württemb, 3:164-167. Trinajstic K. (1999a). Scale morphology of the Late Devonian R Development Core Team (2015). R: A language and environment palaeoniscoid Moythomasia durgaringa Gardiner and Bartram, for statistical computing. R Foundation for Statistical 1977. Alcheringa: An Australasian Journal of Palaeontology, Computing, Vienna, Austria. URL https://www.R-project.org/. 23: 9-19. 228 Bollettino della Società Paleontologica Italiana, 54 (3), 2015

Trinajstic K. (1999b). Scales of palaeoniscoid fishes (Osteichthyes: Xu G.H. & Wu F.X. (2012). A deep-bodied ginglymodian fish Actinopterygii) from the Late Devonian of Western Australia. from the Middle Triassic of eastern Yunnan Province, China Records of the Western Australian Museum, supplement 57: and the phylogeny of lower neopterygians. Chinese Science 93-106. Bulletin, 56: 1-8. Webb P.W., Hardy D.H. & Mehl V.L. (1992). The effect of armored Zelditch M.L., Swiderski D.L., Sheets H.D. & Fink W.L. (2004). skin on the swimming of longnose gar, Lepisosteus osseus. Geometric Morphometrics for Biologist: A primer. 443 pp. Canadian Journal of Zoology, 70: 1173-1179. Elsevier Academic Press, Amsterdam. Wen W., Zhang Q.Y., Zhou C.Y., Huang J.Y., Chen Z.Q. & Benton M.J. (2012). A new genus of basal actinopterygian fish from the Anisian (Middle Triassic) of Luoping, Yunnan Province, Manuscript received 22 August 2015 Southwest China. Acta Palaeontologica Polonica, 57: 149-160. Revised manuscript accepted 16 December 2015 White E.I. (1932). On a new Triassic fish from North-east Published online 30 December 2015 Madagascar. Journal of Natural History, 10: 80-83. Editor Marco Balini